willUms.] DEFECTS IN THE NEW YORK SURVEY. 61 



fossil remains alone meet these requirements. A continuous rock section 

 furnishes us with the order of sequence of these changes, but a classi- 

 iication of the rocks based upon the age of the fossils must not be 

 hampered by stratigraphic or lithologic limits. The time classification 

 can be built up only gradually by wide study of the fossils, and the 

 nomenclature of the formations must be applied, and applied with pre- 

 cision, before the time limitations can possibly be fixed with precision. 



Besides these defects in the final results of the New YTork survey, 

 there were two imperfections occasioned by lack of evidenc, and others 

 due to false generalization. The Devonian system was scarcely more 

 than recognized by its general fauna— the limits above and below were 

 not determined. The upper limit excluded the Catskill formations 

 which were subsequently placed in the system. An equivalency was 

 supposed to exist in Ohio and Michigan between the Chemung and the 

 rocks now called Waverly belonging in the Carboniferous system. The 

 attempts to correlate with the English models resulted in fixing the 

 limit between the Wenlock and lower Ludlow of Murchison between the 

 Corniferous limestone and Marcellus shale of the New York system. 

 The rocks above this limit were correlated with Murchison's Ludlow 

 group and Phillips's Devonian system. 



The imperfection of this work was mainly due to ignorance of the 

 precise relations existing between the two faunas; and, secondly, to the 

 fact that Phillips's fossils were mainly middle and upper Devonian 

 forms, while the lower Devonian species and the lower Devonian type 

 of deposits were not well understood by the New York geologists. 



It was the comparative study of the fossils, and particularly a more 

 careful discrimination of them and better appreciation of the range of 

 the characters they exhibited, which finally cleared up these imperfec- 

 tions. 



Having perfected a scheme of classification, the next step of progress 

 was the correlation of the formations west of New York with the 

 scheme. This was mainly accomplished during the decade from 1840 

 to 1850. The chief discussions of the subject were published between 

 1842 and 1851. 



James Hall published an article in 1842 * in which an attempt was 

 made to correlate the rocks of the States of Ohio, Indiana, Illinois, part 

 of Michigan, Kentucky, Missouri, Iowa, and Wisconsin, with the rocks 

 of New York State. He classified the basins of the Coal Measures into 

 four groups, as follows : first, that of Pennsylvania and eastern Ohio ; 

 a second extending over portions of Indiana, Illinois, Kentucky, and 

 Tennessee; a third in Missouri, and a fourth in Michigan. He traced 

 the underlying " conglomerate" from Pennsylvania to the Mississippi 

 Eiver. The " Old Red sandstone " was not recognized west of the 

 Genesee River in Allegany County, New York; the Chemung forma- 

 tion, which, he remarked, " Lyell compares with the lower part of the 



1 Notes upon the Geology of the Western States, Am. Jour. Sci., vol. 42, p. 312. 



